AD-AOB5 400 MASSACHUSETTS INST OF TECH CAMBRIDGE DEPT OF MATERIA -ETC F/G 11/6SUBMERGED ARC WELDING OF TITANIUM.(U)

UCSEP 18 G HUNTER, 6 B KENNEY, M RING NOGOIR 77-C IBGRUNC LASSIFIED7 .Eol.EE

36

111l_.25 111111.IB 6

MICROCOPY RESOLUTION TEST CHARTNAI;O#4AL BUREAW 01 STANDARDS-I 963-

SECURITY CLASSIFICATION OF THIS PAGE (We De n.

REPOT DCUMNTATOW A(~READ INSTRU CNSREP RT DOC ME TAT ON PA~i-BEFORE COMPLE1 FO1. REORT NMBER 2. GOVT ACCESSION NO, 3. RECIPIENT'S CATALOG NUMW--

TITLE (and Subtitle) n or QnQMED

Submerged Arc Welding of Tianu Technal epor o. 9?

ji' G./ Huntert'G. KenneybM Ring. B.A Russell N RA U

T~. W a 7 1 -7-- 5691

SPERFORMING ORGANIZATION NAME AND ADDRESS 10. PROGRAM ELEMENT. PROJECT. TASKAREA & WORK UNIT NUMBERSDepartment of Materials Science & Engineering

Massachusetts Institute of TechnologyCambridge, MA 02139

It. CONTROLLING OFFICE NAME AND ADDRESS

Office of Naval Research -

800 N. Quincy, Arlington, VA 22217 L40 pagesT MONITORING AGENCY NME & AD 1ES t rollo Office) IS. SECURITY CLASS. (of this report)

Unclassi fied1I15. DECL ASSI F1CATION/ DOWNGRADING

SCMHEDU LE

Approved for Public Release; Distribution Unlimited

17. DISTRIBUTION STATEMENT (of the abstract entored En Block 20, If different from Report)

III. SUPPLEMENTARY NOTES

III. KEY WORDS (Continue on mrvere old* it necessary and identify by block nuinbor)

Welding, Titanium, Fluxes, Oxygen, Nitrogen

ASS40qAfC ,,(AA.K20. A9 AT(Continue an revereet side It necessay and Identify by block numabor)

0.. Studies of flux shielded welding of titanium indicate that acceptable___ submerged arc weld bead shape may not be attained with pure CaF" flux.

Addition of chlorides to the CaF'flux improves the weld bead s~ape butreduces the deoxidizing potentiai of the flux. Use of cored electrodes

Li containing chloride salts produces the most acceptable weld bead profiles. --Preliminary studies of titanium weld penetration in the presence ofseveral fluoride salts indicate that changes in flux chemistry may be useful

M /N00201-661DD FJAN72, 1473 EDITION OP I NOV 616 OBSOLETE Uncl assi fiedSECURITY CLASSIFICATION Of THIS PAGE (,nDote *two

&q.kj41TV CLASSIFICATION O)F THIS PAGE(Whe, Does Entered)

~. ~ in modifying the weld bead profile.

Economic analysis indicates that SAW of titanium can only be

competitive with gas metal arc welding if the flux cost can be reduced

significantly. A possible method of reducing flux cost may be recyclingof the fused slag.

. ri

2ffM_no'-

AV P~id.oD1 st ser I a

. MP

LL

SECURITY CLASSIFICATION OP THIS PAGgr(meM bat. 5aers

7' V* .6 -0..

SUBMERGED ARC WELDING OF TITANIUM

Technical Report No. 3

August 1, 1978 to July 31, 1979

Submitted to:

Office of Naval Research800 N. Quincy St.

Arlington, VA 22217Attn: Dr. Bruce MacDonald

30 September 1978

By:

G. Hunter, G. B. Kenney, M. Ring, B. A. Russell & T. W. Eagar

Department of Materials Science and EngineeringMassachusetts Institute of Technology

Cambridge, MA 02139

Reproduction in whole or in part is permitted for any purposesof the United States Government. Distribution of this document

is unlimited.

.............. _

SUBMERGED ARC WELDING OF TITANIUM

G. Hunter, G. B. Kenney, M. Ring, B. A. Russell & T. W. Eagar

Department of Materials Science and EngineeringMassachusetts Institute of Technology

Cambridge, MA 02139

ABSTRACT

Studies of flux shielded welding of titanium indicate that

acceptable submerged arc weld bead shape may not be attained with

pure CaF2 flux. Addition of chlorides to the CaF2 flux improves

the weld bead shape but reduces the deoxidizing potential of the

flux. Use of cored electrodes containing chloride salts produces

the most acceptable weld bead profiles. Preliminary studies of

titanium weld penetration in the presence of several fluoride

salts indicate that changes in flux chemistry may be useful in

modifying the weld bead profile..

Economic analysis indicates that SAW of titanium can only

be competitive with gas metal arc welding if the flux cost can be

reduced significantly. A possible method of reducing flux cost

may be recycling of the fused slag.

~ .

CONTENTS

ABSTRACT

I. INTRODUCTION 1

II. MATERIALS 4

III. SUBMERGED ARC WELD BEAD APPEARANCE 7

1. Parametric study 7

2. Flux chemistry study 13

3. Conclusions 21

IV. ECONOMIC ANALYSIS OF GMAW vs. SAWOF TITANIUM 23

V. GTAW IN THE PRESENCE OF A FLUX

VI. ADDITIONAL WORK 33

VII. SUMMARY 34

a.- . - . . .

I. Introduction

Welding of heavy section titanium structures in the open

atmosphere presents numerous difficulties, primary of which

is prevention of oxygen and nitrogen contamination of the

weld metal. In order to prevent such contamination, copious

amounts of argon or helium gas, combined with elaborate shields,

are used. These arrangements are capable of reducing the total

contamination level to an increase of less than 100 ppm oxygen

and 20 ppm nitrogen; however, the shields are cumbersome and the

initial joint preparation must be closely controlled to provide

adequate cleanliness. If any errors are made during the

process, the system is unforgiving; and the contaminated weld,

if found, can only be corrected by removal and rewelding. All

of the above requirements lead to increased fabrication costs

for heavy titanium structures.

It has been found that welding of titanium in the presence

of a metal halide flux (e.g. CaF2 ) has the advantage of

providing chemical control of the oxygen content of the weld

metal(1,2) Such control would make the welding process

more forgiving of operator error and hence might be expected

to simplify and reduce the costs of titanium structural

fabrication. Nonetheless, the use of fluxes when welding

titanium presents several additional difficulties; viz.:

-2

1. The fluxes are often hygroscopic and must not

be stored in the open atmosphere.

2. The pure CaF2 flux produces welding arc

instabilities, a somewhat rough weld bead surface

and an unacceptably high weld bead contact angle.

3. In order to control weld metal oxygen content,

the fluxes must be of high purity. The purity

requirement increases the cost of the flux

substantially.

In accordance with these difficulties, the previous

year's effort has been devoted to studying:

1. Use of flux cored electrodes as a means of

reducing moisture contamination of the flux,

2. Methods of improving weld surface smoothness

and weld bead contact angle, and

3. Economic comparison of existing titanium gas-

metal arc welding (GMAW) technology with possible

submerged arc welding (SAW) technology.

" -

. - - -i ... . .. .. __ .. .. __--__-._

Partial success has been achieved in each of the first

two areas, while the economic analysis indicates that SAW can

compete with GMAW only if the flux costs can be reduced

significantly. Possibilities for flux cost reduction are

outlined.

In addition, a preliminary study of the effects of

several metal fluorides on weld penetration has been made.

The results suggest that variation of titanium weld flux

chemistry may provide a -ea-s of controlling weld bead

shape and arc melting efficiency.

Ii

I-4-

II. Materials

All submerged arc welds were made on 19 mm (0.75 inch)

Ti-6Al-4V ELI (extra-low interstitial) plate received from

RMI Titanium Company of Niles, Ohio. The solid welding

electrodes were also Ti-6A1-4V, purchased in three diameters

from Astro Metallurgical Corporation of Wooster, Ohio. The

mill certifications of these materials are given in Table 1.

Flux cored welding electrodes were fabricated from

3.2 mm diameter Ti-3A1-2.5V tubing with wall thicknesses of

0.4 mm and 0.8 mm. The flux, ground to a fine powder, was

tamped into the tube with a long rod. Sections of electrode

approximately six feet long could be made in this manner,

which was sufficient to form one weld bead. Attempts to

fabricate longer sections of electrode by swaging 6.4 mm

or 9.6 mm diameter tubing, failed due to early cracking of

the alloy tube walls. It has been reported that swaging

may be successful if the tubing is pure titanium rather than

an alloy.(3 )

Optical purity CaF2 was the primary flux used throughout

this investigation. Lump crystals were obtained from

Optovac, Inc. of North Brookfield, Massachusetts. Additions

of KCl and NaCl were also used. The KCl crystals were of

optical quality, again produced by Optovac. The NaCl

crystals were reagent grade, obtained from Baker Chemical

1I

-5-

Company. The use of reagent grade NaCi in place of optical

grade was justified on the decreased hygroscopicity of NaCi

compared to KCl.

The fluoride fluxes used in the gas-tungsten arc (GTA)

penetration study were reagent quality powders. No attempt

was made to dehydrate each of these powders as studies of

weld metal chemistry were not to be made. Care was taken

to ensure that each powder remained as dry as possible.

This was practical in every case except KF which was found to

be deliquescent and had to be eliminated from the series.

Table 1. Starting Materials,Base Metal and Electrodes

Base Metal

Ingot Analysis No. 803175

C Fe Al V N 0 H Ti

Certified .02 .17 6.1 3.8 .012 .117 .010 bal.

Check 1 -- -- .0088 .100

Check 2 -- .0121 .176

Annealed 1/2 hour, 927*C, A.C. + 1 hour, 7600C, A.C.

Mechanical Properties

Yield Strength Tensile Strength ElongationMPa (ksi) MPa (ksi) %

Longitudinal 863 (125) 934 (135) 13.5

Transverse 925 (134) 993 (144) 12.5

Welding Electrodes

Diameter, mm (in) C Fe Al V N 0 H Ti

3.2 (.125) .019 .15 6.28 4.11 .006 .078 .002 bal.

Check 2 . -- . . .010 .180 .. .

2.4 (.093) .019 .15 6.28 4.11 .008 .078 .002 ---

Check 2 - --- .-- .010 .180 ... ...

1.6 (.062) .019 .15 6.28 4.11 .008 .078 .009 ---

-7-

III. Submerged Arc Weld Bead Appearance

-As noted in the introduction, pure CaF2 flux has bene-

ficial effects on titanium weld metal oxygen content, but

produces unacceptably poor weld beads, primarily in terms

of weld bead contact angle. Since the bead appearance

observed with pure CaF2 would preclude use of titanium SAW

without further improvement, a study was made of the effects

of process parameters and flux chemistry on weld bead

appearance.

1. Parametric Study

Initially, the effects of weld voltage, current, and

electrode diameter were evaluated in a 23 factorial experiment.

The weld travel speed was held constant at 30 cm/sec. The

dependent parameters consisted of arc stability (monitored as

described in reference 2), weld metal oxygen content, nitrogen

content, weld bead shape, weld metal hardness and bead surface

roughness. The contact angle, penetration, bead height and

width are defined in Figure 1. The results of this series are

shown in Table II.

As can be seen in Table II, the oxygen content of the

weld metal is essentially constant due to the refining action

of the CaF 2 flux. The nitrogen content varies by a large amount,

...... , . .. - .. -

-8-

02BASE BASE

METAL METAL

H : Bead Height

P : Depth of Penetration

W: Bead Width

e): Contact Angle-(e, e 2 )/2

Figure 1. Parameters used to define weld bead shape.

1111FIN Oil n1 -. L, gvwjobwoif i

on t LO 40 a .4 0.

-4.4

40 Nq 4

4) 44D

H (d

43 :4

So a

4jCJ

4

H) M

0A m

U65 4. q 414 0 . f q E

a

440 -0

141

ono

011

t4-4.9-4~ ~~ 65. . . . . . .

011, 0 0 00a 0 10 0 0 14

on .0 In %

- 10 -

Yith the contamination increasing as the electrode melting

rate decreases. This indicates that much of the nitrogen

contamination is due to absorption of atmospheric gases on

the hot unshielded electrode as it emerces from the weld

torch contact tip. This is consistent with previous studies

of nitrogen contamination during titanium SAW. (2) Supplementary

argon shielding would be expected to reduce this nitrogen

contamination.

The arc stability is increased with higher arc voltage.

The bead shape parameters, height, width and penetration

generally increase with increased weld heat input, as expected.

The results indicate that the weld bead contact angle can vary

widely depending upon operating parameters; however, neither

welds 4 nor 6, which had more favorable contact angles, were

acceptable. The cross section of each revealed lack of fusion

between the bead top and the base plate, as shown in Figure 2.

Based upon the above series of tests, weld 3, made using

300 amperes at 36 volts, was chosen as possessing the best set

of operating parameters. The cross section of this bead is

shown in Figure 3. It should be noted that none of these

welds produced a bead which would be acceptable for producing

groove welds in thick plate butt joints. Although improve-

ments in bead shape were made by altering the weld process

variables, no weld made with pure CaF2 flux provided a bead

shape which would eliminate slag entrapment or lack of fusion

defects in multiple pass weldments.

I,

Figure 2. Cross section of weld 6 of Table II. Although

this weld possessed a favorable contact angle, the penetration

of the bead near the edges was deemed inadequate. This weld is

similar in shape to weld 4 of Table II.

A.

B.

Figure 3. A.) Cross section of weld 3 of Table II.B.) Surface of same weld.

This weld was chosen as the standard from which

subsequent improvements in weld bead shape are judged. Note

that the height and contact angle of this bead make it

impractical for producing multipass groove welds.

- 13 -

Two additional process modifications were examined in

hopes of improving the weld bead shape in the presence of pure

CaF2 flux. In the first, the electrode was given a tilt of

either +150 (leading) or -150 (trailing). In the second

modification, welds were made with 3.2 mm diameter tubes,

instead of solid electrodes. The purpose of the latter

modification was to test the changes in electrode melting

behavior (and bead shape) which might be expected when using

an alloy core or flux core electrode. Table III indicates

that both of these modifications result in significant changes

in the weld bead shape, although neither produced the weld

shape which is desired. A leading electrode tilt increases

the weld penetration depth but does not have much effect on

the bead contact angle. Use of tubular electrodes produces

much shallower penetration but improved weld contact angles.

Photos of each type of weld are shown in Figures 4 and 5.

2. Flux Chemistry Study

An alternative to process modification of weld bead shape

is modification of bead shape by changes in flux chemistry.

Investigation of this possibility is of particular interest

as Soviet investigators have long stated that addition of

chlorides to flouride improves their operating characteristics.

If one of these improvements includes bead shape, then use of

the very hygroscopic chlorides might well be justified.

77"

-4 4

£1

.. . . ... ... .......

A.

B.

Figure 4. A.) Cross section of weld 9 of Table III, made

with +150 (leading) electrode tilt.

B.) Rough weld surface of the same weld. Note

that slag is attached to the toe of the weld.

B . I... ..........

20 30 40 50

Figure 5.

A) Cross section of weld 11 of Table III, made with tubular

electrode. Note the much shallower weld penetration when

using tubular electrodes.

B) Surface of same weld bead. The lower contact angle facilitates

slag removal at the toe of the weld, although some slag still

remains. The surface has more uniform ripples but a more

variable width than previous welds.

cc 0 1 1

n In

r. I.- ml - 4 e

iU0 C0! 1 -

0-A0' 1.4

14 E IW in 0 0a tn

0%0

41 04

4J ~ 4;

ON In

*t %0 in

r-40 4) A

.1. O54J (0 ~

,-4 -4 -4 4

0 4

4J.. 4J *4 V. 044 x j 0 O~l 0$4a tT : 0 0 0 I

0 0 0 00 r0.4 n 0~ (l U) Ifl

04 4I N44

U))4.401 00i )w . r * %a 4

4-44 4) 0 V 0 0 0 144 r-" C0 C; 0c 0; -4

*-4H On-.r4 f, r

00

0pr4 0 0 0 0 0.

-4 I 4 r0.4 1* +

.04rq S 4VA #A V

- 17 -

Three methods of adding chlorides to the CaF 2 flux were

tested, viz.:

1. Blended additions of NaCl or KCI, i.e.

mechanical mixture of the powders,

2. Fused additions of KC1, i.e. co-melting of

the salts, and

3. NaCl and KC1 flux cored electrodes operating

in a pure CaF2 granular flux.

The latter technique of adding chlorides through a

tubular electrode has the distinct advantage of keeping the

strongly hygroscopic chloride flux sealed within the cored

electrode. If operable, this technique of chloride addition

should have significant advantage over the use of blended or

fused fluoride-chloride granular fluxes.

The results of the chloride flux tests are shown in

Table IV. The effects of the chloride additions are numerous,

although certain major trends are apparent.

The oxygen content of the weld metal is elevated compared

with pure CaF2 (cf. Table II). This may be due to slight

moisture pick-up due to the chlorides, although a more probable

explanation would be that the increased NaCl or KCl content

has decreased the thermodynamic activity of the CaF Based

upon a Temkin regular solution model, 20 mole percent NaCl or

KC1 in CaF2 would decrease the activity of the CaF2 by 40 to

50%. This could easily account for the reduced deoxidizing

V . . . . . . . . . . . . .

- 18 -

potential of the CaF2. Use of CaCI 2 instead of NaCI or KCl

may restore the deoxidizing potential to the flux.

Unfortunately, CaCI 2 is not preferred due to its high

hygroscopicity.

The nitrogen contamination of the weld metal is markedly

reduced by the chloride additions. This may be due to the

strong fume generation around the electrode caused by the

chlorides. This fume surrounds the hot electrode, thus

reducing atmospheric absorption. The same reduction in

nitrogen contamination may be achieved by external argon

shielding of the electrode.(2)

As in the empty tubular electrode welds described in

Table III, the flux-core tubular electrode welds generally

had elevated oxygen and nitrogen levels. This may be due to

enhanced atmospheric absorption on the electrode, which runs

hotter at the same electrode extension due to the increased

current density. Supplementary argon shielding may reduce

this contamination.

The surface roughness of the chloride flux welds was not

significantly improved over the pure CaF2 flux welds. Although

the ripples were somewhat more uniform, the chloride fluxes

seemed to produce more discoloration and weld spatter than

pure CaF2.

The weld beads are much wider with chloride additions to

the flux. This is likely due to the addition of a new type

-- . . . .- -- ____.. .... ... ... ....... ..... , ..... ___- " -__........... -_I - J A

A. :*i

B) Surface of weld 24. Although the contact angle is improved,

the surface roughness is worse than a weld made with pure

CaF 2 flux.

-20-

of negative ion to the arc plasma. The pure CaF2 flux has

only fluorine as a negative ion. Fluorine has the highest

affinity for electrons of all elements and can be quite

effective in quenching plasmas, i.e. in reducing the electron

concentration of the plasma. SF6 gas is used as an insulator

in high voltage equipment for this very reason. Recent Soviet

studies show that 2.5% SF6 in argon constricts a welding

plasma by a factor of four. (5) Steel welding fluxes, which

are high in CaF2, have been found to narrow the width of

welding arcs. The substitution of chlorides for fluorides in

the present case has obviously widened the submerged arc plasma.

The height of the fused region remains approximately

the same, although the depth of penetration is greatly

enhanced, especially when using tubular electrodes. The

greater penetration and weld bead width indicates that

chloride additions to the flux greatly enhance the melting

efficiency of the process. This is seen in Figure 6, which

shows a much larger weld bead compared with Figure 3.

One of the most significant advantages of using

chlorides in the flux is an improvement in the weld bead

contact angle. Nearly all of the beads made using chloride

additions possessed contact angles which would be acceptable

for multipass groove welds.

The arc stability of the welds made with chloride fluxes

appeared to be similar to welds made with pure CaF 2.

.. . ...i-' ' -

- 21 -

The failure of welds 19 and 20 to follow the common

trends found in the other welds of Table IV may be due to

loss of chloride upon fusion. Only two melts were fused in

the vacuum furnace due to contamination of the system with

chloride salts. The flux compositions shown for welds 19

and 20 in Table IV cannot be taken as accurate due to

evaporative losses during fusion. It is possible that flux

19 had more chloride than flux 20. In any case, the fused

fluxes were not found to behave differently than the blended

fluxes in terms of the parameters studied here. A more

complete evaluation of fused chloride-flouride fluxes is

needed in order to determine if fused fluxes are preferable to

blended fluxes.

3. Conclusions

The weld bead shape may be modified by the choice of

weld process parameters, although no bead made with pure CaF2

flux could be expected to produce defect-free groove welds.

The addition of chlorides to the flux improves the bead shape

and melting efficiency but has a detrimental effect on the

surface roughness and oxygen content of the weld metal. Groove

welds of acceptable shape can be made with chloride containing

fluxes. Fume generation during SAW with chloride containing

fluxes is excessive and should receive further study. Flux

cored welding electrodes produced the least amount of welding

fume.

-22-

14 $m0 W. A. en w 0 m0 a

0 In f" >*0f" C a 0 -WCd 9: IV 0) C: r 4S If

A.qj

4*40

05411 f5 in 05 r- w -W C4 -lW en dln

o44 to In C4 in4 in -0 in in -W. .

00

41 t- in 1Sin% - r 4 C

.o H1 .4 £14 0d4. .13 14 N N

*

xa 41 A6.A A 910A 0 v 41 41 . r

.0 00 .4

0n go 4_4 400. sd . -i .4

.m 0n In 14 eq .4 .4 .4 Cd C"

4' ,4

44rO ~ I1 01'44 ~04

4-44' a04 to a 3

(d 0) 3t 0d 31 f54 w4 N..I.40 0 0 04O) L . .0 4 c

44 N

C1 14 . 4 . 4 .4 .

rI u4 v4 5. v4 -

A~~ ~ ~ ~ v 0004 ud 0 U0 l

o a" 0 o s 0 0 0 10 40 A oa 0 0 0. U'

14~~~~1 C%. e In e b

- 23 -

Future improvements to the process might include:

1. Use of pulsed welding power supplies to improve

metal transfer and bead surface appearance, and

2. Use of Ca - CaF2 fluxes with chloride additions.

22The free Ca which dissolves in CaF 2 may return the

deoxidizing capacity of the chloride containing flux

to the 1000 ppm level. Initial attempts to make

such a flux have been unsuccessful due to inadequate

melting facilities, but in principal these fluxes should

not be difficult to produce.

IV. Economic Analysis of GMAW

vs. SAW of Titanium

The development of a technologically successful method

of submerged arc welding titanium will be of little use unless

the process is also economical. The technological require-

ment of high purity starting flux material has shed some

doubt as to the ultimate economics of SAW titanium. A flux

which costs $18/kg. and is consumed at the rate of 4.5 times

the electrode deposition rate (by weight) can add significantly

to the fabrication costs. Hence, it was deemed necessary bU

- 24 -

to investigate the economic impact of SAW titanium compared

with the current technology of gas metal arc (GMA) welding.

The economic model was chosen to be as simple as possible

yet still differentiate between the relative costs of the two

processes. In many cases large scale production costs are

not available and may only be estimated. The major variables

of concern are electrode cost as a function of electrode

diameter, and the bulk price of high purity CaF2 flux. The

analysis shows the total cost to be most sensitive to these

two factors, which are also the least well known.

The model assumes that capital costs are similar for

both processes and that the processes may be compared based

upon the costs of consumables and labor only. The joint

geometry of each type of weld is also considered to be

constant, although variable joint geometry can easily be

handled by the program. The input data included:

1. Electrode feed rate vs. welding current

2. Weld travel speel

3. Weld voltage

4. Electrode price as a function of wire diameter

5. Rate of flux consumption per ampere of

welding current

6. Argon use rate

-25-

7. Argon cost

8. Labor cost

9. Arc time

The standard conditions for each weld are shown in Table V.

Parameters for GMA welding were obtained from Mr. J. Crisci of

the Naval Ship Research and Development Laboratory. SAW para-

meters were developed experimentally. The standard flux cost

of $18/kg. was estimated as a reasonable bulk price for reagent

quality powder. This figure is substantially less than current

market price for optical quality lump CaF2 in less than 10 kg.

lots, but it is not far from the market price of reagent grade

CaF2 powder in 50 kg. lots. It is assumed that the large lot

savings would approximately equal the melting and purification

costs necessary to refine reagent grade powder.

The results of the comparison based upon the standard

welding conditions is shown in Figure 7. It will be noted that

shielding and electrode costs dominate the total weld cost. For

the GMA process, the single most important cost is the electrode

while in SAW the electrode cost and the flux cost are approxi-

mately equal. Although the larger diameter electrode used in

SAW is somewhat less expensive than the GMAW electrode, the

savings in electrode cost is not sufficient to offset the cost

of the welding flux. It should be noted that the use of auxiliary

argon shielding with the SAW process adds less than one percent

to the cost of the submerged arc weld.

- 26 -

Table V. Standard Welding ParametersUsed in the Economic Analysis of Welding Ti-6A1-4V

GMAW SAW

Welding current, amperes 350 370Welding voltage, volts 20 36Electric power cost, $/KWH - 0.04 -Argon cost, $/m3 - 5.30 -Electrode diameter, mm 1.6 3.2Electrode price, $/kg. 77.38 69.77Travel speed, cm./sec. 0.83 0.67

Labor cost, $/hr. - 15.0 -Arc time, min./hr. 30 24

Shielding gas, I/sec.

A) Torch 0.31 (0.31

optional)B) Trailing

0.71Plate thickness, cm. 5

$500

[FUX8/W $496

$ 4006

CLi

i 300 $283

0" 46%,

o 6%0 CA

(0

04

uJJ zi (D) 0

0 W 4

Figure 7 -- Estimated cost of welding 5 cm thick titaniumplate by GMA or SAW based upon conditions inTable V.

. ......... _ _ _ _

- 28 -

The above analysis assumed that the titanium flux is

discarded for zero scrap value after use. In fact this flux

is relatively inert and it may be possible to reuse the flux

with only minor reprocessing. If the flux cost could be reduced

to $2.25/kg. by multiple use, the SAW process would become

competitive with the current GMAW process. (See Figure 8).

The crossover where GMAW and SAW are equal in cost occurs at

a flux cost of $1.26/kg..

From the foregoing analysis, it is clear that SAW of

titanium is very sensitive to flux cost and electrode cost while

GMAW is primarily sensitive to electrode cost. Although GMAW uses

a smaller diameter electrode, which is slightly more expensive,

the sensitivity to electrode costs between the two processes does

not differ appreciably. From this analysis it can be seen that

the only way in which SAW can compete economically with GMAW

is if the flux cost can be reduced significantly. As noted

previously, this may be possible through recycling of the flux.

Such studies are underway.

$500 -

SFLUX $25/kqW$400 1

WrLIJJ

U) $ 300 -n p300 $296

$284.92 00$200 -

0

$100 - I ol) 10%

W Tf) 61 12.

0 C .E.N.pmr ...

I-I

0D 0 0

C.a)

Figure 8 Estimated cost of welding 5 cm thick titaniumplate by GMA or SAW based upon a flux cost of$2.25/kg and all other conditions as shown InTable V.

- 30 -

V. GTAW in the Presence of a Fluoride Flux

Soviet investigators have reported a process termed "semi-

submerged arc" welding wherein a gas-tungsten arc weld is made

in the presence of a thin paste flux. (2) This process was

originally applied to titanium weldments but it is reported

to be applicable to other metals as well. The flux is

generally composed of a mixture of alkali and alkaline earth

fluorides. It is reported that the flux changes the

efficiency of melting and improves the arc stability. A brief

study was performed to verify these reports.

Fine powders of LiF, NaF, RbF, CsF, MgF 2, CaF 2, SrF 2 and

BaF2 were mixed with propanol and spread across the surface of

a titanium plate in thicknesses of 0.12, 0.25 and 0.5 mm and GTA

welded at 16 volts, 200 amperes, and 25 cm/min., with 0.63 I/sec.

Ar shielding gas. The dependent parameters studied included

arc length, arc width at the midplane, arc stability and weld

shape. The results are shown in Table VI. Although changes

in weld pool shape and size are noted, they are not as

dramatic as has been claimed by the Soviet investigators.

This may be due to the fact that 200 amperes was used in this

study, while Soviet studies may have involved much higher

welding currents. The effects of changes in plasma jet

behavior and weld pool surface tension on weld penetration and

-31-

Table VI. Results of GTA Welding of Titaniumin the Presence of a Paste Flux

Flux Arc Arc Weld WeldThickness, Length, Width, Width, Depth,

Flux mm mm mm mm mm

None -- 28 10 11 2

LiF .12 26 10 12 <1.25 19 9 12 <1.50 14 8 11 3

NaF .12 21 9 5 2.25 26 12 7 2.50 19 10 6 2

RbF .12 25 9 10 2.25 26 11 8 2.50 26 11 6 2

CsF .12 24 11 8 2.25 24 10 9 2.50 23 10 8 2

MgF 2 .12 20 10 6 3.25 21 10 6 3.50 21 10 5 2

CaF .12 27 14 10 2.25 27 13 11 2

.50 26 13 10 2

SrF2 .12 27 10 9 1.25 27 10 9 2.50 27 10 9 2

BaF 2 .12 28 1i 10 1.25 29 12 10 1.50 29 12 11 1

Li27~Ji2

- 32 -

shape should be more pronounced at higher welding currents.

Further studies at varying current levels are planned, although

a welding chamber will be required to contain the fumes as the

operator experienced respiratory irritation during this study.

I _ _ _ _ -..

- 33 -

VI. Additional Work

In addition to the studies reported, work has proceeded

in two other areas during the past year, viz. construction of

a chamber for controlled atmosphere testing by SAW, GMAW and

GTAW, and construction of a Cl - HCl gas train for use in2

drying the hygroscopic chloride fluxes. Although each of

these are now operational, no data has yet been generated as

a result of their construction. Each is expected to be of

importance in studies during the coming year.

- 34 -

VII. Summary

An experimental and economic study of submerged arc

welding of titanium has shown the following:

1. Modifications to the welding process can influence

the weld bead shape when using pure CaF2 flux, but

the changes are insufficient to provide the control

of bead shape which is required for production of

groove welds.

2. Modification of the flux chemistry by the addition of

chlorides improves the melting efficiency and provides

acceptable bead shape, however, the deoxiding potential

of the flux is reduced.

3. Economic comparison of GMAW and SAW of titanium

indicates that SAW will be economic only if flux

recycling is possible.

4. Gas tungsten arc welding in the presence of a fluoride

flux does affect the arc behavior and melting efficiency.

The extent of the change depends upon the flux composition.

• ., _.r. .. : . : ._ ... ,. ,. .... , . . . ,4-

-35-

References

1. S.M. Gurevich, V.N. Zamkov, S.D. Zagrebenajuk and

N.A. Kushnirenko, "The Effects of Fluxes Containing Rare

Earth Elements on the Structure and Properties of Welds

in VT15 Alloy," Avt. Svarka, No. 4, 1964, pp. 93-94.

2. C.S. Chai, J.J. Gullotti and T.W. Eagar, "Submerged Arc

Welding of Titanium," Technical Report No. 2, submitted

to the Office of Naval Research, Contract N00014-77-C-0569,

30 September, 1978.

3. J. Crisci, David Taylor Naval Ship Research and Development

Center, Annapolis, private communication, September 1979.

4. F.D. Richardson, Physical Chemistry of Melts in Metallurgy,

Vol. 1, Academic Press, New York, 1974.

5. B.N. Bad'yanov, "Change in the Dimensions of the Arc

Electrical Conductivity Zone When Gaseous Halides Are

Introduced into Argon," Avt. Svarka, 1977, No. 4, pp. 67-68.

• .. ..... ... ... ' ..... :- .. i , . . . ... ...... i: _.. , .. . . ~> 2 " "

[LL!LLII7

~wr,